It sounds like an idea from the Matrix: learn about any topic in just a few minutes, but it is a reality and it is the AnyTopic application, a website that is opening a new path for online learning, allowing users to create personalized mini audiobooks for free.
Users enter a topic or question and the platform finds the most suitable online content: news, blogs, existing podcasts and more. The app then creates a mini audiobook from this material, lasting between 3 and 25 minutes.
These short, personalized episodes are designed to fit into users’ busy lives, Perfect for listening to on your commute or relaxing at nightThe app’s personalization features tailor each audiobook to the user’s interests and learning pace, providing personalized content for each individual. When AnyTopic started, it focused on finding the most innovative solutions linked to learning.
“We were investigating the most advanced models for voice generation, with the goal of using one of these models to generate audiobooks “We have a lot of custom audiobooks,” explains Sam Wesley, co-founder of AnyTopic and one of the developers of Amazon’s Alexa. “Of course, an audiobook can’t exist without a narrator, but using traditional voice talents wasn’t an option.”
Thus, they needed an AI voice generator that was not robotic but expressive. They had to keep users engaged, as if they were listening to a great teacher. But they also needed software that supported high-quality audio. They knew that poor audio quality would make users much less likely to want to keep listening.
For this they chose to use ElevenLabs, a company specialized in developing natural voice softwarefounded in 2022 by Piotr Dąbkowski, a former Google machine learning engineer. Of course, selecting the model was just the first step; they then had to integrate it with their app. Fortunately, this was seamless. “We were able to easily incorporate the ElevenLabs Speech API as the final step in our established audiobook generation process,” Wesley adds.
ElevenLabs’ real-time latency generates audio content almost instantlymeaning students can dive into a topic without long wait times. Plus, API integration means users can handle everything in one easy-to-use interface, making learning fast and simple on the go.
“We are excited about the future of personalized content, particularly in the audio space,” said Daniel Rascon, co-founder and CEO of AnyTopic. “ElevenLabs has helped us realize this vision much faster and at the level of quality we strive for. We are excited to create more engaging audio experiences for our listeners.”
The operation is very simple. Go to the website, register with an email address and select the topic and language in the box (more than 50, including Spanish, Galician and Catalan). Then select the level (beginner, intermediate or expert) and the length (between 3 and 25 minutes). Finally, click create. After 10 minutes, you will have your audio text.
It may be that We don’t like the voice that much, it still seems somewhat artificial and with a strange cadence, but once we get used to it, it is a very useful option to learn about basically any topic in our free time.